EVALUVATINGROAD TRAFFIC ACCIDENTS USING DATAMININGTECHNOLOGY ABSTRACT:-Roadtraffic safety is an important perturbation for government transportauthorities as well as common people. Road accidents are ambivalent and notable to be predict the incidents. And their survey requires the informationaffecting them. Road accidents cause difficulties which are get bigger at analarming rate. Controlling the traffic accidents on roads is a crucial task. Togive safe driving suggestions, clear and careful study of roadway traffic datais critical to find out the variables that are nearly to fatal accidents.
Increasing the number of vehicles from past few years has put lot of pressureon the existing roads and ultimately resulting in increasing the roadaccidents. A road traffic accident is any harm due to collision originatingfrom, terminating with or involving a vehicle partially or fully on a publicroad.I. INTRODUCTION:- In modern life, accidents havebecome daily happening.
Every day we hear the news of the accident on thetelevision, or through internet .During accident many people die at the spot, someothers may injured very severely. By witnessing an accident one can understandthe horror of it. There are several reasons for road accidents, some of them areincreasing the number of vehicles, careless driving, violating traffic rulesetc. Whenever a road accident occur there are various types of damage takesplace ,which could be in the form of human beings, infrastructure which isdamage to the government and many other administration damages . Poor roadway maintenancealso contributes accidents.
But still many people continue to neglect andignore the danger involved in the accidents. In this paper we are analyzingsome methods and algorithms to find out the problems occur in road accidents.Section1 elucidate literature survey,Section 2 elucidate conclusion. LITERATURE SURVEYThe paper 1 describes the associationrule mining, its classifications and the atmospheric components like roadwaysurface, climate, and light condition do not strongly influence the fatal accidentrate. But the human factors like being alcoholic or not, and the impact havestrongly affect on the fatal accident rate. Acommon mechanism to recognize the relations between the data stored in hugedatabase and plays a very significant role in repeated object set mining isassociation rule mining algorithm. A classical association rule mining methodis the Apriori algorithm whose main aim is to identify repeated object sets toanalyze the roadway traffic data.
Classification in data mining methodology focusat building a classifier model from a training data set that is used toclassify records of unrevealed class labels. The Naïve Bayes technique is oneof the probability-based methods for classification and is based on the Bayes’hypothesis with the probability of self-rule between every set of variables. The author applies statisticsanalysis and Fatal Accident Reporting System (FARS) to solve this problem.
Fromthe clustering result some regions have larger fatal rate but some others havesmaller. When driving within those risky or dangerous states, people take moreattention. When the task performed, data seems never to be sufficient to make astrong choice.
If non-fatal accident data, weather condition data, mileagedata, and so on are available, more test could be executed thus more advicecould be made from the data.In paper 2, K-modes clusteringtechnique is a framework that is used as an initial work for division ofdifferent road accidents on road network. Then association rule mining are usedto recognize the various situations that are related with the occurrence of anaccident for the entire data set (EDS) and the clusters recognized by K-modesclustering algorithm.
Six clusters (C1toC6) are used based on propertiesaccident type, road type, lightning on road and road feature identified by Kmodes clustering method. On each cluster association rule mining is applied aswell as on EDS to create rules. Powerful methods with higher raise values are takenfor the inspection. Rules for various clusters disclose the situations relatedwith the accidents within that cluster. These rules are compared with the rulescreated for the EDS and resemblance shows that association rules for EDS does notdisclose correct data that can be related with an accident. If more feature arepresented large information can be identified that is associated with anaccident.
To buildup our methodology, we also performed analysis of all clustersand EDS on monthly or hourly basis. The results of analysis assist methodologythat performing clustering prior to analysis helps to identify better anduseful results that cannot obtained without using cluster analysis.The paper 3 performs statisticaland empirical analysis on State Highways and Ordinary District Roads accidentaldatasets. The need of the study is to analyze the traffic accident data of SH’sand ODR’s to assign the black spots and accidental elements, part to controlthe harm caused by the accidents. The basic necessity of the analysis is tocheck the traffic associated dataset through Exploratory VisualizationTechniques, K-means and KNN Algorithms using Rstudio.
. The term accident blackspot in management of road traffic safety defines a place where accidents arebeen focus historically and to analyze the accidental data using exploratoryvisualization techniques and machine learning algorithms. These techniques andalgorithms are used on the traffic accidental dataset to get the desired outputin order to reduce the accident frequency. Exploratory Visualization Technique is atechnique to anatomize and examine the sets of data in order to abridge andencapsulate the important characteristics with visual and pictorial method.Exploratory Visualization analysis can be performed using scatter plot,correlation analysis, barplot, clustered barplot, histogram, pie chart etc.
Machinelearning concentrates on algorithm designing and makes predictions on sets ofdata. It includes Supervised (KNN Algorithm) and Unsupervised learning (K-meansAlgorithm).This paper present result by resembling the above three mining techniques and assigns the causeof accident, accident prone area, analyze the time of accident, examine thecause of accident and scrutinize the litigators vehicle.
In paper 4, describesabout a frame work that uses K-mode clustering technique as a primary task fordividing 11574 accidents on road network of Dehradun (India) from 2009 to 2014.Then an association mining rule are used to find out the various context associatedwith instance of an accident for both the whole data set and clusters find out byK-modes clustering algorithm. Then compare the findings from cluster basedanalysis and entire data set. The results shows that the amalgamation of k modeclustering and association mining rule is very encouraging, as it producesimportant facts that would remain hidden if no segmentation has been performedprior to generate association rules. Also a trend analysis has been performedon each clusters and entire data set.
By trend analysis it shows that beforeanalysis, prior segmentation of data is very important. This paper put forwarda frame work based on cluster analysis using k-mode algorithm and associationmining rule. By using cluster analysis as a primary task can group the data intodifferent homogeneous parts. It is the first time that both association andclustering rule are used together to analyze the data’s for road accidents. Theoutput of the study proves that by using cluster analysis as a primary task, itcan help in removing heterogeneity to some extent in the road accident data.
)Based on attributes accident type, road type, lightning on road and road feature,K -modes clustering find six cluster (C1–C6). Association mining rule have beenapplied on each cluster as well as on entire data set to generate rules. Forthis analysis strong rules with high lift values are used.The paper 5 describespurpose of data mining methods in the field of road accident investigation. .
Association rules are used to identify the patterns and rules that aresubjected the cause the occurrence of road accidents. An efficient method for updatingthe index year after year could be designed. Additionally, further analysis oftraffic safety data using data mining techniques are allowed.
Cluster analysis evaluates dataobjects without consulting a common class label. The objects are clustered orarranged on the basis of maximizing the intra class similarity and minimizingthe interclass similarity. Outlier analysis: A database having data objectsthat do not satisfies the general behavior or model of the data. These data objects are also called outliers.Evolution analysis which defines and models consistencies or trends forobjects whose behavior changes over time.
We are currently build up byconsidering several issues, changes in clash occurrence may have someaftereffect for traffic safety measures in certain countries. The determinationof specific precautionary measures to overcome clashes requires study of otherfactors such as the identification of specific road sections that need work,etc..
It analyzed the traffic accident using data mining technique that couldpossibly reduce the fatality rate. Using a road safety database enables toreduce the fatality by implementing road safety programs at local and nationallevels. The paper 6 describes data miningtechniques to analyze high-frequency accident locations and further identifydifferent factors that affect road accidents at specifying locations. We firstpartitioned the accident locations into k groups based on their accidentfrequency poll using k-means clustering algorithm. Association rule mining algorithmis used to reveal the correlation between different elements in the accidentdata and understand the characteristics of these locations. Hence, the major significancewill be the evaluation of the outcomes. Data mining has been proven as areliable technique to analyzing road accident data. Several data miningtechniques such as clustering, classification and association rule mining arewidely used in the literature to identify reasons that affect the severity ofroad accidents.
It is the first time that k-means algorithm is used to identify high- andlow-frequency accident locations based on accident count as it provides sometechnical measures to divide the accident locations based on threshold values. The road accident datasetand its analysis using k-means clustering and association rule mining algorithmshows that this approach can be reused on other accident data with moreattributes to identify various other factors associated with road accidents.In paper 7 describestheresults from analysis of traffic accidents on the Finnish roads by applyinglarge scale data mining methods. The set of data collected from road trafficaccidents are vast, multidimensional and diverse. TheFinnish Road Administration between 2004 and 2008 data was collected for thisstudy. This set of data contain more than 83000 accidents and 1203 of which arefatal. The main aim of this is to examine the usability of robustclustering, association and frequent item sets, and visualization methods tothe road traffic accident analysis.
The output shows that the pick out datamining methods are able to produce intelligible patterns from the data,detecting more information that could be increased with more detailed andcomprehensive data sets. Most of the fatal accidents occur due tothe condition of single roadway main roads outside built-up areas where thepermitted speed varies typically between 80-100km/h. Aged and young drivers havelarge contribution to the high risk accidents in highways. Most of the surveysreported that one of the major reason for accidents among young people areconsumption of alcohol. From the analysis it is understand that failure ofroads and end user groups are responsible for accidents at certain limit.
Thispaper 8 is to represent a Traffic Accident Report and Analysis System (TARAS)through data mining using Clustering technique. Detect the causes of accidentsis the main aim of this paper. The transport department of government of Indiaproduced the dataset for the study contains traffic accident records of theyear and look into the performance of J48. The classification accuracy on thetest result discloses the three cases such as accident, vehicle and casualty.
GeneticAlgorithms is used for the future selection to lower the measurements of thedataset.. More detailed area specific information from accident locations and circumstancesare needed. With the help of this paper, the analysis can be done and thereforepreventive measures can be taken. It can help the government to keep track ofrecords of the accidents, causes of accident, vehicle number, vehicle owner’sname and address.
. With the current data it is possible to identify the risky roadsegments and the road user groups responsible for accidents in certain environments.The viewer or user can also make their own account for viewing the site .youcan view the data about causality .
Our system will provide the graphical view ofthe accidents with respect to the data entered into the system according to theperiod .This system will provide the solutions as accidents causes. So thatwith the help of this system government can take the necessary actionsaccording accidents cases.1) Accurate Location ofaccident2) GPS integration3) Government IDAuthentication for user Data4) Advanced Filtertechnique Accident Solution prediction.
The paper9 describes application of data mining techniques on road accidents by usingmachine learning algorithms that determines accident rate in the future todecrease clash deaths and wounds. The accident datasetcontains traffic accident report of various cities examined by using machinelearning algorithms to predict the accident rate. Itimplemented hybrid approach that performed with higher accuracy rate ascompared to other methods to be described. The machine learning techniques is usedfor to reduce accidents and saves life. We have to expandthe classification accuracy of road traffic accidents types; data quality hasto be added.
In paper 10 describesabout a method called Innovators Marketplace on Data Jackets. InnovatorsMarketplace on Data Jackets used to externalize the value of data through ally.For analyzing the rate of traffic accidents on urban area methods such as factor analysis, structureequation modeling and data mining are used here.
To construct traffic accidentrisk evaluation model different indexes such as total number of accidentsreported, fatality rate injury rate arecombined. To identify the connection between different factors populationstructure information, vehicle information, road characters are used. In Herewe focused on urban data, applied structural equation modeling to find out the importantfactors associated with traffic accident.
Important factors are population structure,vehicle information, structure of road etc. This paper describes six factors byconstructing an accident risk causal framework based on urban data and the componentfactor sets of each feature and influence on traffic accident.Reference1 Analysisof Road Traffic Fatal Accidents Using Data Mining TechniquesLiling Li,Sharad Shrestha, Gongzhu Hu2 Analysing road accident data using associationrule miningSachin Kumar; Durga Toshniwa3Black Spot and Accidental Attributes Identification on State Highways andOrdinary District Roads Using Data Mining Techniques.Gagandeep KaurDepartment of Computer Engineering Punjabi University Patiala Punjab, IndiaHarpreet Kaur Department of Computer Engineering Punjabi University PatialaPunjab, India8Traffic Accident Report Analysis using DataMining TechniquesMrs. Kanchan Gawande1 Ambikesh Pandey9 A Radical Approachto Forecast the Road AccidentUsing Data Mining TechniqueAnupama Makkar ,HarpreetSingh Gill